The term infective endocarditis properly includes the whole world of microorganisms that can cause the disease. Fungi, rickettsiae (Coxiella burnetii, which causes Q fever), Chlamydia sp, and perhaps even viruses have been implicated in endocarditis, although bacteria are still the predominant cause. Substantial advances in the isolation of microorganisms and improvements in serologic testing and molecular detection have widened the spectrum of causative organisms. Uncommon species of streptococci, emerging pathogens such as Bartonella sp and Tropheryma whipplei, the increase in fungal pathogens among nosocomial cases and immunocompromised patients, and increasing resistance among “typical” endocarditis pathogens such as enterococci present unique diagnostic or therapeutic challenges [5,6,16,19,20,21,22,25].

Table 80.1 summarizes the most common pathogens from large series of endocarditis cases occurring since 1985 [1,2,3,4,6,7,8,9,14,26,27,28,29]. Those series with a large number of injection drug users [1,6,14,18,27], patients on chronic hemodialysis [1,2,17,22], transplant recipients [19,20], and those series reporting healthcare-associated IE [1,6,7,16,22,23,25] tend to report more cases caused by S. aureus. Viridans streptococci occur more frequently but no longer predominate among noninjection drug user populations and the elderly [9]. Identification to species level among the viridans streptococci may have important therapeutic and prognostic implications. The Streptococcus anginosus (milleri) groups (S. anginosus, S. constellatus, and S. intermedius) are frequently associated with abscess formation and tend to cause severe disease but cause endocarditis less often than other viridans streptococci [30,31].
S. anginosus is the least likely among the three species to cause abscesses and the most likely to be associated with endocarditis [31,32]. The nutritionally deficient streptococci include the following genus and species: Abiotrophia defectiva, Granulicatella adiacens, Granulicatella para-adiacens, Granulicatella balaenopterae, and Granulicatella elegans [33,34]. Together these organisms constitute 3% to 5% of cases of endocarditis caused by viridans streptococci [34]. These organisms require pyridoxal, the active form of vitamin B₆, for growth. Unlike other species of viridans streptococci, these organisms are tolerant to penicillin, and at least one series [33] has found decreased susceptibility to penicillin, extended spectrum cephalosporins, and macrolide antibiotics. High relapse rates are described especially when patients are treated with penicillin alone [34]. Most cases of viridans streptococcal endocarditis (80%) are caused by Streptococcus sanguis, Streptococcus mitis, or Streptococcus mutans [14,18,29,30]. There do not appear to be any statistically significant differences in the symptoms, demographics, or complications among patients with infections caused by this group of organisms. Newer species of viridans streptococci continue to be described.

Enterococci rank third in frequency of isolation in most series, including healthcare-associated cases and those among patients on hemodialysis. Among the non–viridans streptococci, pneumococci are still relatively uncommon causes of endocarditis (1% to 3% of all cases). The proportion of cases caused by β-hemolytic streptococci has not increased since 1980; infections with group B and group G are seen most frequently [32,35,36]. Patients with these infections usually have underlying valvular disease, numerous predisposing factors, most notably diabetes mellitus, and acute onset of their infection [32,35,36]. Streptococcus bovis deserves mention for several reasons. First, this organism group has undergone extensive reclassification based upon DNA–DNA reassociation studies, a description of which is beyond the scope of this chapter [37]. Second, some of the newly described species and subspecies (S. gallolyticus spp gallolyticus) are more frequently associated with endocarditis and with benign and malignant disorders of the gastrointestinal tract, while others are more frequently associated with meningitis (S. gallolyticus spp pasteurianus). In some series, this organism group has been increasing in frequency in Europe and South America [9], particularly among the elderly and among patients with chronic liver disease [4,37,38,39]. When this organism is isolated, the patient should be carefully evaluated for gastrointestinal tract malignancy, although it may occur months to years after the bacteremic episode [38]. Since these species and subspecies are difficult to differentiate using traditional microbiological methods, most clinical laboratories will likely continue to call them S. bovis or nonenterococcal Group D streptococci.

S. aureus has increased in frequency and accounts for more than 50% of cases in more recent series [1,2,6,9,16,17,19,25,27]. In several recent prospective studies, including data from the International Collaboration on Endocarditis-Prospective Cohort Study (ICE-PCS), patients with S. aureus endocarditis were more likely than patients with IE due to other pathogens to have a shorter duration of symptoms before diagnosis, to be hemodialysis dependent, and to have other serious comorbidities such as diabetes mellitus or other chronic illnesses [9,40,41,42,43]. Patients with S. aureus IE were also more likely to have severe sepsis with persistent bacteremia, major neurologic events, systemic embolization, and death than patients with IE caused by other bacteria [41,42,43]. In these studies, patients with S. aureus IE frequently had healthcare-associated or nosocomial acquisition and were more likely to have MRSA infection than patients with community-acquired S. aureus. In the ICE-PCS series [38], a multivariate model identified the following patient characteristics associated with MRSA IE: persistent bacteremia, chronic immunosuppressive therapy, intravascular devices as sources, and diabetes mellitus. Overall, MRSA now accounts for 25% to 50% of the cases attributable to S. aureus [2,25,40,41,42,43]. Persistent bacteremia correlated with infection caused by MRSA, and risk of embolic phenomena was negatively associated with oxacillin resistance [40,41,42,43].

Coagulase-negative staphylococci (CoNS) are recognized pathogens on prosthetic valves and close to 8% of cases on native valves are now caused by these organisms [44]. The majority of the CoNS species recovered are Staphylococcus epidermidis [44]. Staphylococcus lugdunensis has emerged as a particularly aggressive pathogen that causes a destructive native valve endocarditis, frequently following vasectomy or other procedures involving breaks in the skin in the perineal area [45]. In spite of universal susceptibility to β-lactams and other agents, mortality attributable to this pathogen is high, possibly related to the large vegetations frequently seen with this organism, leading to valvular dehiscence, abscess formation, and systemic embolization [45].

Before 1980, endocarditis caused by Gram-negative organisms comprised less than 3% of cases. Recent series report that Gram-negative organisms now account for 4% to 10% of all native valve endocarditis [3,6,15,46], but these rates vary by geographic location, whether the infection is community or healthcare associated and the type of Gram-negative pathogen involved. Within this subset is the HACEK group (Haemophilus sp, Aggregatibacter [previously Actinobacillus] actinomycetemcomitans, Cardiobacterium hominis, Eikenella corrodens, Kingella kingae), which accounts for 2% to 5% of cases [8,9,26,46]. A. actinomycetemcomitans is the HACEK species most frequently involved in IE [46]. HACEK organisms are fastidious, nonmotile, slow-growing coccobacilli that require a mean of 3.3 days of incubation in automated blood culture systems for growth [46]. The HACEK organisms rarely cause endocarditis in patients without preexisting valvular disease or in the absence of predisposing factors [46]. Non-HACEK Gram-negative endocarditis (Enterobacteriaceae and others) remains relatively rare and is seen primarily among debilitated patients with healthcare-associated infections related to medical devices or surgery [21,47,48].

Among the emerging pathogens of the 1990s are Bartonella sp. This genus continues to expand [49,50]. Seven species and subspecies, namely B. quintana, B. henselae (the agent of cat scratch disease), B. elizabethae, B. vinsonii spp berkhoffii, B. vinsonii spp arupensis, B. koehlerae, and B. alsatica have been implicated in cases of endocarditis [49,50]. B. quintana, the agent of trench fever, has been reported to infect middle aged, homeless male alcoholics without known underlying valvular disease. Contact with animals is a frequent association, and ectoparasites such as scabies, lice, and fleas are proposed as possible vectors of disease [49,50,51]. The majority of patients with B. henselae endocarditis have a previous history of underlying valvular disease and report contact with cats [49,50,51]. Characteristically, patients with Bartonella endocarditis present with a subacute course and large vegetations [49,50,51]. Because of the fastidious nature of the organism and serologic cross-reactivity between antibodies to B. quintana and Chlamydia sp, it is likely that cases of Bartonella endocarditis constitute a proportion of cases previously diagnosed as culture negative or due to Chlamydia sp [49,50,51]. Currently, approximately 3% of cases of endocarditis are secondary to Bartonella sp [51].

In spite of improvements in blood culture and serological techniques, negative blood cultures can occur in up to 31% of cases [1,2,3,7,52]. There are several reasons cited for negative blood cultures in IE: (a) prior antibiotic administration; (b) infection with fastidious, slow-growing organisms (e.g., Bartonella sp, fungi, Chlamydia and Coxiella spp); (c) infection with nonbacterial organisms such as fungi; and (d) endocarditis in patients with an indwelling cardiac device such as a
pacemaker [52,53]. Two recent surveys [52,53] of culture-negative cases in France over two decades used serological studies and molecular methods to augment blood cultures in determining the etiology for more than 348 [52] and 740 [53] patients, respectively. In the initial study of definite cases of IE, in the 79% of patients in whom an etiologic agent was determined, C. burnetii and Bartonella sp predominated, accounting for 76% of the total [52]. Other rare bacteria included T. whipplei, Mycoplasma hominis, various streptococci, and Legionella pneumophila [52]. Twenty-one percent did not have an etiology determined of whom 79% had received prior administration of antibiotics [52]. In the more recent prospective series, both definitive and possible cases were included and an etiologic diagnosis was determined in 64.6%. The same group of organisms predominated [53].

IE among injection drug-users has increased in the new millennium and S. aureus is by far the most common cause [54,55]. Enterococci, enteric Gram-negative bacilli, Pseudomonas, Candida, and other yeasts are also important [6,9,54]. Polymicrobial endocarditis is more common among injection drug users than in noninjection drug users [6,9,18].

The common causes of early-onset PVE are CoNS (mostly S. epidermidis), S. aureus, enterococci, diphtheroids, Gram-negative bacilli, and fungi. Among the fungi, Candida sp are most common and have emerged as causes of both early- and late-onset disease especially among patients with healthcare-associated infection [56]. However, late-onset disease is still caused mainly by organisms such as CoNS and streptococci, although S. aureus accounts for about 11% of cases [9,18,57,58]. This difference is thought to be explained by intraoperative or early postoperative contamination of the prosthesis with resistant hospital flora in early PVE. Late cases represent either smoldering infection with relatively avirulent organisms seeded at the original surgery or subsequent transient bacteremias, such as those that induce endocarditis on native valves [58].

The laboratory model of endocarditis is a rabbit in which a catheter passed through a valve produces mild trauma with the elaboration of a fibrin–platelet thrombus. Subsequent injection of bacteria either through the catheter or at a distant vascular site leads to infection of the traumatized valve [59]. It appears that the fibrin–platelet thrombus allows for avid binding of the bacteria [6]. Adherent bacteria induce blood monocytes to produce cytokines that contribute to further enlargement of the vegetation [60]. As the vegetation matures, the bacteria become fully enveloped, which allows for persistence by avoiding host defenses.

This model conforms to the propensity of damaged human valves toward endocarditis. Transient bacteremia with mouth flora, predominantly viridans streptococci, during chewing, tooth brushing, and the like explains the pattern observed in subacute bacterial endocarditis [6]. More virulent organisms such as S. aureus seem to be able to invade even normal hearts. There are several factors expressed by this pathogen that make it more virulent. In addition to surface fibronectin-binding proteins that facilitate adherence, S. aureus produces exoenzymes and exotoxins that are controlled by global regulators, such as accessory gene regulator (agr) and staphylococcal accessory regulator (sar), the expression of which permit tissue invasion and destruction [6]. Intravenous drug users combine the injection of contaminated materials with particulate and often irritant matter, probably accounting for the frequency of endocarditis in this setting and the propensity for right-heart involvement [54,61,62]. The use of intravascular central lines reaching near the tricuspid valve or even crossing tricuspid and pulmonic valves reproduces the rabbit model of endocarditis in humans. The introduction of bacteria through these lines causes the specter of iatrogenic endocarditis.

Once the fibrin–platelet thrombus has become infected, the pathologic process is the enlargement of this mass into a vegetation and invasion of tissue by the infection with eventual disruption. In addition to the mass of the vegetation, there are perforations or total erosions of valve cusps, rupture of chordae tendineae, fistulas from the sinus of Valsalva to atrium or pericardium, and burrowing myocardial abscesses.

Depending on the valve involved, the physiologic consequences may be predicted. Rarely, a vegetation will be so large as to function as an occlusive or stenotic lesion [62]. More often, the tissue destruction process predominates and valvular incompetence results. New regurgitant murmurs of mitral, tricuspid, or aortic origin may acutely stress the heart with resultant congestive failure. Aortic valve disease carries the worst prognosis [62] for several reasons: (a) the heart tolerates acute aortic insufficiency least well; (b) pericardial tamponade or massive left-to-right shunt may develop if a sinus of Valsalva aneurysm erodes into the pericardium or right atrium, respectively; (c) heart block may occur if a myocardial abscess invades the conducting system; and (d) aortic valve ring vegetations are most likely to be flipped into the coronary arteries, infarcting already overworked muscle. These catastrophes are all even more likely in the presence of a prosthetic aortic valve, in which case the infection has its seat at the annulus. Tricuspid valve endocarditis is the most benign. Even total tricuspid insufficiency can be tolerated for a time, and acute right-side heart failure is not as life threatening as is the pulmonary edema of left-sided failure.

The vegetations themselves may break off in whole or part as emboli to the brain, viscera (spleen and kidney are particularly common targets), coronary arteries, and notably in fungal endocarditis, large arteries of the extremities. Septic emboli to the lungs can result in pulmonary infiltrates, often nodular and sometimes cavitating. Emboli to other organs produce infarction, which is usually bland, although splenic abscess, brain abscess, and even purulent meningitis may occur in staphylococcal endocarditis. The most common cerebral lesion, however, is embolic infarct with the clinical appearance of a stroke [62,63]. The smaller vascular lesions of endocarditis may be of an immunologic, vasculitic nature or truly embolic and suppurative in character. Emboli to the vasa vasorum or vasculitis of the arteries lead to mycotic aneurysms of both cerebral and peripheral vessels. The cerebral aneurysms are generally asymptomatic until they rupture and present as subarachnoid or intracerebral hemorrhage. Peripheral mycotic aneurysms may come to attention because of their obvious enlargement and frequent overlying inflammation. Other phenomena that fall into this category are the cutaneous stigmata of endocarditis—Osler’s nodes, Janeway lesions, splinter hemorrhages, and petechiae—as well as the frequent renal involvement. Kidney pathology may take several forms: localized renal infarcts, vasculitic glomerulonephritis, acute diffuse glomerulonephritis thought to represent immune complex disease, renal cortical necrosis, and interstitial nephritis likely related to antibiotic administration [62].

Endocarditis is diagnosed on the basis of signs and symptoms that reflect the pathology: fever, embolic phenomena, and evidence of valvular dysfunction. A continuous bacteremia is characteristic and, indeed, highly suggestive of endovascular infection, although the entity of culture-negative endocarditis also exists. The frequency of various findings in IE is shown in Table 80.2 [1,3,4,9,14,18,27].